Does receiving an eating disorder diagnosis increase the risk of a subsequent alcohol use disorder? A Danish nationwide register-based cohort study.

Abstract

No large-scale, longitudinal clinical study has examined whether patients with different types of eating disorders (ED) have an increased risk of a subsequent alcohol use disorder (AUD). This study aimed to assess the ongoing risk of receiving a diagnosis of AUD following a first-time diagnosis of anorexia nervosa (AN), bulimia nervosa (BN), or unspecified ED (USED). Retrospective cohort study. Danish nationwide registries, January 1994 to December 2018. A total of 20 759 ED patients and 83 036 controls were followed from the date of first ED diagnosis (index date) until the date of first AUD diagnosis, death, emigration, or the end of the study. Controls were selected in a 1:4 ratio and matched on month and year of birth, gender and ethnicity. We obtained data on ED (AN, BN, USED; exposure) and AUD (abuse/dependence; outcome) diagnoses as well as sociodemographics and other psychiatric diagnoses. Time to AUD was generated from the index date. Risk of AUD after the index date was assessed among those without a prior AUD diagnosis while adjusting for sociodemographics and prior psychiatric diagnoses. Compared with controls, an increased relative risk of AUD after the index date was observed in AN patients throughout the study lasting 15+years (adjusted hazard ratios [HRs] ranging from 2.49 [99% CI = 1.46, 4.25] to 6.83 [2.84, 16.41]), in BN patients during the first year of follow-up and from 2years onward (2.72 [1.66, 4.44] to 17.44 [6.01, 50.63]), and in USED patients during the first year and 2-15years of follow-up (2.52 [1.54, 4.14] to 14.17 [5.86, 34.27]). In all three groups, estimates were highest during the first year, particularly among BN patients. Patients with anorexia nervosa, bulimia nervosa, or unspecified eating disorders appear to have an increased ongoing risk of receiving a diagnosis of alcohol use disorder following their first eating disorder diagnosis compared with controls.